Assembling the Weather Station

Assembling the Weather Station#

I spent some time today assembling one of the weather station carrier boards I recently received from PCBWay.

I placed the order just after midnight on 26-03-2025, and it arrived on the afternoon of 04-04-2025, giving a turnaround of roughly 9 days to Northern Ireland. For a professionally manufactured PCB, that’s a very reasonable lead time and something worth factoring into future iterations.

The boards arrived vacuum-sealed and surrounded by protective packaging. Straight out of the pack, they looked clean — good silkscreen alignment, no obvious defects, and a consistent finish across the batch.

At this stage, though, appearance doesn’t count for much. The real test is how well the board assembles and whether the design holds up in practice.

Assembling the SMD Components#

I decided to assemble the SMD components first, using the larger hotplate.

To begin with, I preheated the PCB to around 50°C. This helps the Sn42Bi58 solder paste adhere to the pads more easily without prematurely reflowing.

I then applied the solder paste using a syringe, aiming for a consistent amount on each pad.

With the paste applied, I placed the components using a pair of tweezers. For inspection and alignment, I used a jeweller’s loupe, which made it much easier to verify placement — particularly on the smaller passive components.

Once everything looked aligned, I gradually increased the temperature to around 145°C, allowing the paste to begin activating, and then continued up to approximately 175°C until full reflow occurred. At that point, the solder joints transitioned to a smooth, silver finish — a clear indicator that the process had completed successfully.

I then removed the board from the hotplate and allowed it to cool naturally.

What I Would Do Differently#

Overall, the board assembled cleanly and the solder reflow behaved as expected. However, there are a couple of things I would adjust on the next pass.

First, I would reduce the preheat temperature slightly to around 40–45°C when applying the solder paste. At 50°C, the paste was slightly too fluid, which made precise application more difficult. A lower temperature should still allow the paste to adhere while improving control.

Secondly, I would change the order in which solder paste is applied — specifically leaving finer-pitch pads until later in the process. During this build, a couple of these pads began to dry out before reflow.

Applying additional flux resolved the issue, but it highlights an opportunity to improve the workflow.

There was also an issue with the USB connector, which I’ll cover shortly.

For the next iteration, I will likely follow this process:

• Apply solder paste to larger pads (typically passives)
• Place passive and larger components first
• Apply solder paste to finer-pitch pads (ICs and connectors)
• Place connectors and ICs
• Perform reflow and allow the board to cool

This build was also slightly slower than expected, as some components were still in unclear packaging, requiring me to double-check part identification during assembly.

Testing the Board Part 1#

Before applying power, I performed a basic set of checks using a multimeter in continuity mode.

This involved working methodically around the board to confirm expected connections and identify any unintended shorts.

During this process, I discovered an issue — a short to ground.

The USB Connector#

I had a suspicion this part might cause problems — and it did.

I was slightly careless when applying the solder paste, relying a bit too much on the idea that the connector would “snap into place” during reflow. In this case, it didn’t. The result was a short on one of the fine-pitch pins — specifically, a short between 5V and ground.

With the fault identified, I decided to remove the connector and rework it properly using hot air.

To do this, I first protected nearby components using Kapton tape to shield them from the heat. I then preheated the board on the hotplate to around 120°C to reduce thermal shock and make removal easier.

Using a hot air gun set to approximately 350°C, I applied heat to the connector until the solder reflowed and the component could be lifted cleanly from the board.

After removing the connector, I allowed the board to cool before carefully reapplying solder paste — this time taking more care with both the amount and placement.

I then returned the PCB to the hotplate (again at around 120°C) and used the hot air gun to reflow the connector into place.

After repeating the continuity checks, the short was gone and all connections appeared correct.

Continuity Testing Approach#

At this stage, I follow a simple but structured approach when checking for faults:

• Check connectors first
  Any short at a connector — particularly before protection circuitry — can propagate issues into other systems, so this is always the first point of inspection.

• Check power rails
  Verify there are no shorts between power and ground, both before and after any protection circuitry.

• Check expected connections
  Confirm that key nets are continuous between components and that signals route as intended.

• Check for unexpected connections Confirm that adjacent pads or connections haven’t accidentally been connected due to a reflow issue.

This process is deliberately methodical. It helps catch critical issues early, before any power is applied to the board.

Powering It Up#

With the initial assembly complete, the following key components were in place:

• Power LED and resistor
• Main USB connector
• Power OR-ing diodes
• LDO regulator and associated capacitors
• I²C pull-up resistors
• USB power negotiation resistors

At this point, the board had passed continuity checks and no obvious shorts were present.

Rather than connecting it directly to a USB power source, I opted to use a bench power supply with current limiting enabled. This provides a controlled environment and reduces the risk of damaging components if an issue is still present.

I initially set a low current limit and gradually applied power while monitoring the board for any unexpected behaviour, such as excessive current draw or components heating up.

Once power was applied, the board behaved as expected. Both the 5V and 3.3V rails were present and stable, with no signs of excessive current consumption, overheating, or instability.

At this stage, the power subsystem appears to be functioning correctly.

Attaching the ESP-C3-SuperMini#

The next step was to solder the ESP32-C3 SuperMini module along with its associated decoupling capacitor.

This followed a similar process to the earlier SMD assembly, although this time I used the smaller hotplate to apply more localised heat to the board.

I began by preheating the area to around 50°C to assist with solder paste application. The ESP32-C3 SuperMini was placed, and the temperature was increased to approximately 145°C to initiate reflow, before being raised to around 175°C to complete the process.

Once reflow was complete, the board was removed from the hotplate and allowed to cool naturally.

After assembly, I repeated the continuity checks — focusing in particular on the power rails and the module connections. I verified that:

• The GPIO pins were correctly connected to the header pins
• The SDA and SCL lines were properly routed to the relevant connectors
• No shorts were present between power and ground

With these checks complete, the board was ready for the next stage of testing.

Second Power Up#

With the ESP32-C3 SuperMini soldered in place and continuity checks complete, I reapplied power to the board.

On power-up, the module started successfully and the preloaded LED blink test program began running.

There were no signs of excessive current draw, overheating, or instability — indicating that both the power subsystem and module integration were functioning correctly.

What I’d Change in Revision 2#

So now that the board is up and running, what would I change?

• The power LED is slightly too close to the USB connector. I would move it approximately 0.25 inches further away, as proximity isn’t required and future rework of the connector may be necessary.
• Improve the external power connector labelling by clearly marking polarity. A simple ‘+’ indicator would significantly reduce the risk of incorrect connection.
• Refine the USB connector footprint and include dedicated test points for power and ground. This would simplify debugging and make it easier to identify faults such as shorts during bring-up.
• Replace the pull-up resistor jumper pads with proper header pins to make the board more flexible for reuse as a general-purpose breakout.

Next Steps#

Now that the board appears to be functioning correctly, the next step is to attach the header pins and connect the BME280 and OLED display to verify that these components can be successfully initialised.

If you are working on custom carrier boards, sensor integration, or ESP32 prototypes and need an experienced engineering partner, let’s discuss your project at Protoworks (NI).